Analysis of Surface Reaction of LiMnxFe1-XPO4 /Li4Ti5O12 Cell during High Temperature Cycle

Abstract

Introduction Energy storage has a key role in balancing to the grid and providing emergency back-up. Li-ion batteries containing olivine metal phosphates (LiMPO4, M = Mn, Fe) as cathode and spinel lithium titanate (Li4Ti5O12: LTO) as anode are promising solutions for energy storage[1,2]. Because of LiMPO4 has a rigid poly-anion structure and Li4Ti5O12 has “zero strain” characteristic during charge-discharge cycling, LiMPO4/Li4Ti5O12 cell shows high safety and cyclability over wide temperature region. In particular, Mn-rich olivine cathodes (LiMnxFe1-xPO4: LMFP) have been paid attentions for stationary applications due to their high energy density. Moreover, high energy density and stability in high temperature region enable to exclude cooling systems and downsize cell modules. However, there are a few reports about high temperature performance of LMFP/ LTO cell[1]. This research is aimed to reveal cyclability of LMFP/LTO cell under high temperature condition including surface and impedance analysis and show the potential for stationary power applications. Experimental Laminated LMFP/LTO cells using 1.2M LiPF6 in 1:2 ethylene carbonate (PC) / diethyl carbonate (DEC) electrolyte with 500mAh capacity were used for electrochemical measurements. The charge-discharge tests were carried out between 2.7 and 1.5V with 2C rate at 60°C. XPS analysis was performed to analyze electrode surfaces. Results and discussions Fig. 1 shows discharge curves and cycling performance of LMFP/LTO cell under high temperature condition of 60°C. LMFP/LTO cell showed stable capacity retention over 95.9% after 500 cycles. Considering high capacity retention after 500 cycles, it was suggested that degradation of electrode materials were significantly small. Fig. 2 shows Cole-Cole plots of (a) half-cells and (b) LMFP/LTO cell. Corresponding to half-cells measurement, semicircles in low and high frequency region of LMFP/LTO cell are attributed to charge transfer resistances of LMFP cathode and LTO anode, respectively. Semicircle of low frequency expanded, while that of high frequency was almost same during initial 100 cycles, indicating that LMFP resistance increased while that of LTO anode increased little. Fig. 3 shows surface compositions of LMFP and LTO before cycling and after 100 cycles at 60°C estimated from XPS analysis. As for LMFP cathode, the amount of LiF increased after 100 cycles, meaning LiPF6 decomposition was a main side reaction. In addition, surface coverage on LMFP cathode increased after 100 cycles. It resulted in impedance increase and during initial 100 cycles. As for LTO anode, the amount of LiF did not increase and surface coverage was almost same during initial 100 cycles. In the literature, capacity fade has been reported due to LiF formation during cycling test[1]. LiF is a reaction product of PF5 or HF and active Li of LTO. Thus, LiF formation led to capacity fade of the cell. In this research, capacity fade was suppressed, because Li loss from LTO due to LiF formation has not observed during cycling test. This difference possibly results from the amount of H2O impurity or/and surface reactivity of LTO. In this research, LMFP/LTO cell demonstrated excellent capacity retention and small impedance increase at 60°C. According to XPS analysis, surface coverage increased for LMFP cathode after 100 cycles. Cole-Cole plots showed resistance increase of LMFP cathode only during initial 100 cycles, but impedance increase was negligible above 100 cycles. Based on above results, it is suggested that stable SEI formed after 100 cycles and decomposition of electrolyte on LMFP cathode was suppressed above 100 cycles. As for LTO anode, LiF formation was suppressed and Li loss from LTO anode was not occurred during cycling test unlike previous report[1]. These results indicate that stable SEI formation on LMFP cathode and LTO anode enabled to realize excellent cycling performance of LMFP/LTO cell under high temperature conditions. Cycle life performance after a prolonged cycling tests and detailed analysis will be reported. References V. Borgel et al, J. Electrochem. Soc, 160 (4) A650-A657 (2013) R. Castaing et al, J. Power Sources, 744-752, 267 (2014) Figure 1